Lists Home |
Date Index |
"Simon St.Laurent" wrote:
> Dave Winer's written a "Rebuttal to REST":
Rebuttal to "Rebuttal to REST"
This essay claims to rebut some things I've written.
Actually it knocks down some strawmen. Dave admits he didn't
even read the most recent article (the one he linked to). He
just skimmed it. To me is quite disrespectful to "rebut"
something you haven't read. I make technical points in that
article and a rebuttal would address those points.
Here is my rebuttal to the "rebuttal":
Prescod is an advocate of a philosophy called REST, which
suggests that there is a single correct way to expose
XML-based services over the Internet.
I don't think anybody has said that. In fact, the article
cited contains this statement:
Now your application may have different requirements than
Google's so the analysis may be quite different. HTTP is
not necessarily better than SOAP-RPC for every application
in the world. In my experience, however, the HTTP solution
is better for any project where a public (as opposed to
individual or departmental) interface is needed for a
Furthermore, I've never said that REST is the "only way". I've
said that REST is the better way...when compared to RPC, for
services that need to scale, evolve and be integrated with
other services. Most SOAP advocates agree that an XML-centric
view is more scalable and evolvable than an RPC-view.
Unfortunately an interoperable, non-RPC way of using SOAP is
not well articulated in the specification so it is hard to
compare REST to it.
Dave: I'll concede that anything you could do with XML-RPC or
SOAP could also be done with REST.
What I'm waiting for you to either concede or disprove is that
there are some concrete additional benefits that you get from
REST that you do not with RPC. Once you acknowledge this then
we can have a meaningful discussion on when RPC is the best
solution and when REST is. But until you do, we keep going
around in circles every other month. You say: "the RPC version
is as good as the REST version." Then I disprove that. Then
you say: "but in some circumstances the RPC version might be
better" and I acknowledge that. And then we start the
conversation at the same place again the next time. If you
care about technology then let's talk about when one strategy
is a better technological choice then the other.
I apologize for making the debate somewhat personal, Dave. But
every time the REST debate comes up, you jump in without
feeling the need to learn anything new or evolve your
position. Most others can debate the issue and come out on the
other side either with agreement or at least with some new
technical ideas but you seem to resist that. I feel that we
will have the same conversation three months from now. This
lack of progress is very frustrating. I feel it stems not from
a real inability to understand the issues but from your
personal relation to the XML-RPC and SOAP technologies.
Dave says: If Google were to redesign their interface to
please a few people would be a waste of resources.
Google started with a REST+XML interface a year ago and a
REST+HTML interface three or four years ago. Rather than
document or publicize their work of a year ago, they took it
pay-only, then redesigned it to be SOAP-based and did a big
press release. That was a waste both of resources and of
momentum. Furthermore, Google has no idea how many people they
would make happy with an XML-based interface because the XML
community has not had an opportunity to try to build such an
interface into applications. Meerkat arose organically from
the existence of HTTP-fetchable XML. Nobody planned it when
they started putting those XML files on the Web.
Dave says: SOAP is, as Prescod acknowledges, a juggernaut.
It's better, imho, to accept that it's here, above all the
objections that have been raised.
Why would we ever consider technical issues when we could just
go with the flow and follow Microsoft? I guess that the rest
of us can just turn off our brains and wait for Microsoft to
bless technologies. I personally refuse to succumb either to
the oligopoly or "Neilson ratings" theories of technology.
There are real, technical differences and those differences